CORPSE is an extension to the CobraPy environment to make some extra functionality - I initiated this because I missed functions for reconstruction of specific models.

The toolset includes functions for:

• mapping gene expression values to reactions
• get essential genes/rxn depending on the global lower and upper and local thresholds
• calculate distances on FVA results for different samples/simulations of the same model
• extract tissue specific models using fastcore

Simply use pip to install the package:

pip install git+https://github.com/Porthmeus/CORPSE.git

Alternatively clone/download the package and install it manually

git clone https://github.com/Porthmeus/CORPSE.git
cd CORPSE
python setup.py install

Here a simple example of how to use fastcore. First we load the E. coli model from cobrapy as a toy model and we take 5 random reactions as a core set.

import cobra
from corpse import simpleFastcore
import random

eco = cobra.io.load_model("textbook") # load the model
core_eco = random.sample([x.id for x in eco.reactions], 5) # sample 5 reactions

To run fastcore instantiate a fastcore object and call run():

fast_mod = simpleFastcore(model = eco, core_set = core_eco)
core_mod = fast_mod.run()

This will run fastcc and fastcore on the model and return the tissue specific model. One can run also only fastcc or fastcore on the model provided:

# run only fastcc
fast_mod.fastcc()
consist_mod = fast_mod.get_model()

# run only fastcore
fast_mod.fastcore()
core_mod = fast_mod.get_model()

To change the solver for the problems, simply change the solver of the initial model object provided:

eco.solver = "cplex"
fast_mod = simpleFastcore(model = eco, core_set = core_eco)

To define the active reactions, usually one uses omics data to map these to the reactions. corpse offers to map transcriptomic or proteomic data (or any other measure which can be associated to genes/proteins) to the model in questions. Given you have transcriptomic data where the rows contain genes and the columns samples one can use the following code to map.

import corpse
import cobra
import pandas

mod = cobra.io.read_sbml_model("path/to/model.xml")
df = pandas.read_csv("path/to/csv", index_col = 0)
mapper = corpse.omicsMapper()
RAS_df = mapper.mapExpressionToReaction(model = mod, dataframe = df)

Set protein = True if you want to use the gene names instead of the IDs of the model.

Check out the coreSetFinder function - it is an easy way to apply different thresholding strategies like global and local thresholds and eases up the process of finding the right threshold setting for the given project.

Write a good documentation with examples how to use the library